Projects

Table of Contents

Marine Tourism Development Data Analysis

Project Type
AI & Machine Learning

Area/Sector
Marine Tourism and Destination Development

Academic Institution
Dalhousie University

HQP’s
Vanshi Negandhi

Problem: As Atlantic Canada continues to grow as a popular sailing destination, it is challenging to prioritize investments in infrastructure and tourism activities that would best support the needs of visitors and local communities.

Solution: Use AIS tracking data to inform coastal development strategies can help make data-driven decisions.

Data Structure and Data Collection Design

Problem: Data from different sources is inconsistent and does not translate well to stakeholders.

Solution: Strategically collect, organize, clean, and optimize data from internal test engines and customer assets and create visualizations and interactive dashboards for stakeholders.

Project Type
Data processing

Area/Sector
Oil/gas

Academic Institution
Acadia University

HQP’s
Matthew O’Hara

Biomass Identification on vessel (FishInFrame)

Problem: In commercial fisheries monitoring, human reviewers spend hours of manual labour watching video footage from cameras on fishing boats to make sure boats are following the rules and fishing appropriately.

Solution: Develop and deploy a multi-layer neural network using cloud infrastructure to automatically indicate the presence of biomass on active vessels. 

 

Website Graphics (1404 × 332 px) (1)

Project Type
AI and Machine Learning

Area/Sector
Commercial Fisheries

Academic Institution
Dalhousie University

HQP’s
Antor (Kazi) Hasan, Ayush Verma

Zooplankton Computer Vision

Problem: To quantify copepod zooplankton populations, thousands of images from the area of interest are taken. Experts must then analyze each image for population counts. This is time consuming and takes time away from more important tasks.

Solution: A computer vision algorithm is able to analyze the images and recognize which contained zooplankton. The model will place all images with a high likelihood of containing the target organism into another folder. This subset of images allows the experts to rapidly quantify the population while ensuring no occurences were missed.

 

Project Type
Computer Vision

Area/Sector
Commercial Fisheries

Academic Institution
UNB

HQP’s
Mayank Anand

Marine Canary: Mussel Closure Prediction

Problem: With rising water temperatures toxic algae blooms, which is harmful to humans the surrounding ecosystem, are increasing in frequency and severity. 

Solution: Using computer vision to track mussel behaviour, their movements can be used to predict such harmful events before other sensors may detect an issue. 

 

Project Type
Computer Vision

Area/Sector
Marine Research

HQP’s
Amous Qiu

Surface Vessel Object Detection Dataset Development

Problem: An object detection algorithm requires data corresponding to the classes it is trying to detect. For niche environments and applications, such data is not readily available.

Solution: The dataset was manually created by collecting images and videos from various databases, and other sources from the internet. 

 

Project Type
Data Collection

Area/Sector
ROV

HQP’s
Akshit Patel

Drone Movement and Landing Data Analysis

Problem: Since 1964, weather balloons carrying radiosondes (battery-operated devices that transmit meteorological data) have been released twice a data from 31 sites across Canada. Once the weather balloons burst, the radiosondes plummet back to Earth.

Solution: Develop an algorithm that, depending on the surrounding conditions. will instruct a drone which site or alternate collection point to return to. 

 

Project Type
Data Strategy and Analytics

Area/Sector
Weather and Sustainability

HQP

Divyansh Saini

Electric Marine Vessel Data Analysis

Problem: Commercial lobster vessels currently primarily operate using combustible fuel engines, with typical trips involving significant engine idling. To support BlueGrid’s mission of increasing vessel electrification, analysis of the current paradigm would support their hypotheses. 

Solution: Performing statistical analysis of trip data concluded the hypothesis of a strong linear correlation between trip time and energy consumption. 

Project Type
Statistical Analysis 

Area/Sector
Marine Transportation

HQP
Rutvik Joshi

Phase 3: Smart Atlantic Buoy Redundancy Model Deployment

Problem: The Smart Atlantic Buoy situated in the approach to the Halifax Harbour shares critical sea state information to ships entering the harbour. If it goes down, there is no longer a source of that information. 

Solution: By using a nearby Buoy’s similar data, a regression model is able to use it as a proxy for the Halifax Harbour Buoy. 

4

Project Type
Machine Learning Deployment

Area/Sector
Marine Operations

HQP
Dhruv Patel, Amous Qiu, Rohini Chandrala, and Parth Champaneria 

Stella Maris Cybersecurity

Problem: With a 900% increase in cyberattacks on the maritime industry, all platforms are required to meet the IMO (International Maritime Organisation) guidelines. Due to this new platform, there is a need for data governance, data security and overall data accuracy. 

Solution: Building out management systems and security policies using the new platform called Stella Maris, allowing partners and clients to engage with the platform by sharing real-time data. 

Project Type
Data Strategy

Area/Sector
Marine Cybersecurity

HQP
Alyssa Jones

Stella Maris Data Analysis and Visualizations

Problem: Numerous sensors located on Stella Maris are producing constant streams of real-time data, including sensors in the early stages of development.  

Solution: Produced clear and meaningful data visualizations to provide usable analysis of Stella Maris data and provide manufacturers with materials to support sensor development and commercialisation. 

Project Type
Data Visualizations

Area/Sector
Marine Research

HQP
Liam Bennett

Stella Maris Governance and Processes

Problem: Grafana is not fully intuitive to users that are not familiar with it. There is no document or guide where the user can as a reference to find answers or help with any arising issue.  

Solution: A user guide with instructions on how to use Grafana, including some examples of problems that users may encounter, and its respective solutions. 

Project Type
Data Processing

Area/Sector
Marine Research

HQP
Mariana Rincón

Pairing Meteorological and Power Data for Marine Hybrid Electric Boats - Phase II

Problem: The purpose of this project is to test the feasibility of predicting boat characteristics (fuel usage CO2, etc) using data from GlasOcean measuring devices like accelerometers. The research also aims to investigate the relationships between data on external weather conditions and vessel power behaviour for a given set of vessel attributes.

Solution: A torque model was created to predict the torque generation of the boat engine based on the boat’s speed and wave height corresponding to the boat trip. Secondly, a fuel consumption model was created to predict the fuel consumed by the boat engine based on the boat’s speed and wave height over the boat’s journey. 

Project Type
Predictive Model

Area/Sector
Marine Operations/ Energy Efficiency

HQP’s
Karansingh Sudhirsingh Chauhan

Ship Stability

Problem: Vessel stability parameters are important in estimating their under-keel clearance, which is one of the OMC’s main product offerings. Unfortunately, often in ship reports such stability parameters are incorrect. 

Solution: Using a numerical machine learning model, we are able to use vessel characteristic data to infer ship stability parameters. This model can assess if current values are correct, otherwise model outputs may be used as more representative values to reality. 

Project Type
Data Classification

Area/Sector
Marine Transportation

HQP’s
Abdelrahman Abdelrahman

GIT Fleet Dashboard

Problem: Internal report writing was repetitive, and required too much time and effort from the data science team. The same insights are required for each report pertaining to individual ships in their fleet. 

Solution: A network pipeline was created to automatically push new data to a database, which is then periodically queried by the dashboard backend to automatically update the dashboard and generate reports. 

Project Type
Dashboard Creation

Area/Sector
Marine Transportation

HQP’s
Saumya Shah

Ship Stability

Follow us on social media